562 THE VOYAGE OF H.M.S. CHALLENGER. 



Even Archidiscus pyloniscus may possibly represent the same form among the 

 Porodiscida as Triodiscus lenticula among the Pylodiscida ; this important form 

 indicates clearly the close affinity of the two families. 



If we take the latter, nearly identical form as the common starting point of both 

 families of Cyclodiscaria, then probably Triopyle and Triolene must be regarded as 

 retrograde forms, derived from Triodiscus in the one case, from Archidiscus in the 

 other, by reduction of three interradial arm-chambers, whilst three perradial only 

 remain. But it is also possible that the most simple form, Triolene, originated 

 independently from some Cenosphcera, three simple radial chambers, like the latticed 

 central chamber, being derived from the latter by apposition in three equidistant radii, 

 whilst three other radii between them remained free. In this case the other genera of 

 Pylodiscida are derived from their common ancestral form Triolene. 



Adopting this latter view, we find that all eight genera of Pylodiscida, here dis- 

 tinguished, may be regarded as following members of a continuous series. If the 

 three simple arm-chambers of Triolene, surrounding the equal central chamber, become 

 united at their distal ends by a concentric equatorial ring, then originates Triopyle ; 

 and this graduates into Triodiscus by fenestration of the three open gates between the 

 three latticed arms. Whilst these three genera form together the subfamily Triopylida, 

 a second family, Hexapylida, is composed of three other analogous genera, in which the 

 same process of development becomes repeated. 



Pylolena, the most simple form of Hexapylida, arises from Triodiscus by the 

 development of three new arm-chambers (of the second order) which are apposed at the 

 distal end of the three primary arm-chambers (of the first order) in the same radius. If 

 the distal ends of these three secondary arm-chambers become united by a concentric 

 latticed ring or girdle, we get Hexapyle (with six open gates, two in each radius), and 

 if its six gates become afterwards closed by loose lattice-work, we arrive at Pylodiscus 

 (a repetition of Triodiscus). 



A third subfamily, Discopylida, is formed by the building of a chambered equatorial 

 girdle around the margin of Pylodiscus. This girdle has quite the same structure as 

 the similar chambered rings or girdles of the Porodiscida and Coccodiscida. . Between 

 the two sieve-plates of the disk surface is enclosed a variable number (twelve to twenty- 

 four or more) of chambers, imperfectly separated by radial beams, which connect the 

 margin of the Pylodiscus-shell with an outer peripheral concentric ring. In Disco- 

 zonium this marginal ring is perfect, whilst in Discopyle it is interrupted by a peculiar 

 large opening, a " marginal osculum " surrounded by a corona of spines, quite the same 

 remarkable formation which we encountered in Ommatodiscus among the Porodiscida. 



All Pylodiscida are therefore triradial (with three perradial arms and three interradial 

 gates between them), and many of them have a great resemblance to certain triradial 

 Porodiscida and Spongodiseida, perhaps not only a morphological resemblance, but also 



